Learning Semantically Enhanced Feature for Fine-Grained Image Classification

TL;DR

This paper introduces a simple, efficient method for fine-grained image classification that enhances feature semantics without complex localization modules, achieving competitive results with minimal additional parameters.

Contribution

The method learns semantically enhanced sub-features by channel grouping and regularization, providing a plug-and-play module that improves fine-grained classification with only image-level supervision.

Findings

Achieves comparable performance to state-of-the-art methods.

Parameter-efficient and easy to integrate into existing models.

Validated effectiveness through extensive experiments.

Abstract

We aim to provide a computationally cheap yet effective approach for fine-grained image classification (FGIC) in this letter. Unlike previous methods that rely on complex part localization modules, our approach learns fine-grained features by enhancing the semantics of sub-features of a global feature. Specifically, we first achieve the sub-feature semantic by arranging feature channels of a CNN into different groups through channel permutation. Meanwhile, to enhance the discriminability of sub-features, the groups are guided to be activated on object parts with strong discriminability by a weighted combination regularization. Our approach is parameter parsimonious and can be easily integrated into the backbone model as a plug-and-play module for end-to-end training with only image-level supervision. Experiments verified the effectiveness of our approach and validated its comparable performance to the state-of-the-art methods. Code is available at https://github.com/cswluo/SEF